This study presents a theoretical discussion analyzing and explaining the use of group methods in feeding back diagnostic data to organizations. A new design -- the peer group-intergroup model -- is presented and compared to the traditional family group model. Data evaluating one implementation of this design showed that it was associated with senior officers of a bank changing their attitudes toward the relevance of a diagnostic study and their willingness to consider changing their own behavior. Changes in organizational practice also followed the feedback. An explanatory model for the new design, derived from the general theoretical considerations, was supported by data taken from the feedback sessions. (Author)